Mechanical cavopulmonary assist device cage for Fontan patients

Project description

provide you with an outline to write the paper based on the outline, you should be very specific and write in details about each point and explain it further.. I will upload some very helpful sources. you should be writing 3 pages on the design specification and 2 pages on the social and environmental impact. the outline have the outline for the whole research to give an insight of the project. please read through the outline first and use the sources I uploaded and then write the research in detail with full explanation and understanding of the project.
Mechanical Cavopulmonary Assistance of a Patient-Specific
Fontan Physiology: Numerical Simulations, Lumped
Parameter Modeling, and Suction Experiments
*Amy L. Throckmorton, *James P. Carr, *Sharjeel A. Tahir, *Ryan Tate, *Emily A. Downs,
*Sonya S. Bhavsar, †Yi Wu, ‡John D. Grizzard, and §William B. Moskowitz
*Department of Mechanical Engineering, School of Engineering, Virginia Commonwealth University; ‡Department of
Radiology, School of Medicine, Virginia Commonwealth University; §Division of Pediatric Cardiology, Children’s Hospital
of Richmond and School of Medicine, Virginia Commonwealth University, Richmond, VA; and †Department of Mechanical
Engineering, School of Engineering, Behrend College, The Pennsylvania State University at Erie, Erie, PA, USA
Abstract: This study investigated the performance of a
magnetically levitated, intravascular axial flow blood
pump for mechanical circulatory support of the thousands
of Fontan patients in desperate need of a therapeutic
alternative. Four models of the extracardiac, total cavopulmonary
connection (TCPC) Fontan configuration were
evaluated to formulate numerical predictions: an idealized
TCPC, a patient-specific TCPC per magnetic resonance
imaging data, and each of these two models having a
blood pump in the inferior vena cava (IVC). A lumped
parameter model of the Fontan physiology was used to
specify boundary conditions. Pressure-flow characteristics,
energy gain calculations, scalar stress levels, and blood
damage estimations were executed for each model.
Suction limitation experiments using the Sylgard elastomer
tubing were also conducted. The pump produced
pressures of 1–16 mm Hg for 2000–6000 rpm and flow
rates of 0.5–4.5 L/min. The pump inlet or IVC pressure
was found to decrease at higher rotational speeds.
Maximum scalar stress estimations were 3 Pa for the
nonpump models and 290 Pa for the pump-supported
cases. The blood residence times for the pump-supported
cases were shorter (0.9 s) as compared with the nonsupported
configurations (2.5 s). However, the blood damage
indices were higher (1.5%) for the anatomic model with
pump support. The pump successfully augmented pressure
in the TCPC junction and increased the hydraulic energy
of the TCPC as a function of flow rate and rotational
speed. The suction experiments revealed minimal deformation
(<3%) at 9000 rpm. The findings of this study
support the continued design and development of this
blood pump. Key Words: Single ventricle physiology—
Cavopulmonary assist device—Fontan physiology—Blood
pump—Artificial right ventricle—Pediatric circulatory
support—Mechanical cavopulmonary assist—Computational
fluid dynamics—Total cavopulmonary connection.
Introduced in 1968, the Fontan procedure was initially
created to treat patients with tricuspid atresia
(1,2). Since then, it has been generally employed in
the palliative treatment of patients with complex congenital
cardiac malformations having only one functional
ventricle (1–3). Although many surgical
variations of this procedure have been investigated
over the years, the total cavopulmonary connection
(TCPC) is the most broadly used Fontan procedure
and is performed in a staged approach (1). Surgical
advances of this procedure have led to more streamlined
vessel configurations, improved hydraulic efficiency,
and fewer sutures as compared with earlier
Fontans. Even though these advances have improved
surgical outcomes, a long-term consequence of the
Fontan physiology is cardiovascular decompensation
to congestive heart failure due to a progressive
increase in pulmonary vascular resistance and
preload limitation (1,3). Heart transplantation is a
doi:10.1111/j.1525-1594.2011.01339.x
Received February 2011; revised April 2011.
Address correspondence and reprint requests to Dr. Amy
Throckmorton, Department of Mechanical Engineering, Virginia
Commonwealth University, 401West Main Street, Rm. E3221, PO
Box 843015, Richmond, VA 23284, USA. E-mail: althrock@vcu.
edu
Presented in part at the 7th International Conference on Pediatric
Mechanical Circulatory Support Systems and Pediatric Cardiopulmonary
Perfusion held May 5–7, 2011, in Philadelphia, PA,
USA.
Artificial Organs
••(••):••–••,Wiley Periodicals, Inc.
© 2011, Copyright the Authors
Artificial Organs © 2011, International Center for Artificial Organs and Transplantation andWiley Periodicals, Inc.
1
viable treatment option for these patients if they can
survive the donor organ waiting period. In conjunction
with surgical optimization of the TCPC, the
introduction of periodic or short-term cavopulmonary
assist may also serve as a possible long-term,
clinical management strategy for Fontan patients (3).
We are developing a collapsible, percutaneously
inserted, magnetically levitated axial flow blood
pump as a mechanical cavopulmonary assist device
for the Fontan circulation (4–6). Thousands of firstgeneration
Fontan recipients and more recent surgical
patients are being routinely treated for symptoms
of heart failure; despite an incidence of only two in
1000 births in the USA, these patients utilize healthcare
resources disproportionate to their numbers
with costs exceeding $500 million per year (3). This
pump is intended to serve as a bridge-to-transplant,
bridge-to-hemodynamic stability or bridge-tosurgical
reconstruction for Fontan patients. It is
designed to provide 4 weeks of temporary mechanical
circulatory support with replacement possible for
prolonged support. Implantation of the pump
involves insertion into a femoral vein and advancement
toward the inferior vena cava (IVC) or extracardiac
conduit at the entrance of theTCPC junction.
This study presents the numerical analysis of our
blood pump mechanically supporting a reconstructed,
patient-specific, anatomical model of the
Fontan physiology.To the authors’ knowledge, this is
the first study to present modeling of mechanical
cavopulmonary assistance using a patient-specific
Fontan physiology. The numerical analyses included
the analysis of pressure-flow characteristics, energy
assessment calculations, and a blood damage assessment
(7).We examined hydraulic performance conditions
for five pulmonary arterial pressures (10, 14,
18, 22, and 26 mm Hg), a range of pump rotational
speeds as dependent upon the model under evaluation,
and flow rates ranging from 0.5 to 4.5 L/min. In
addition, a lumped parameter model of the cardiovascular
system for a Fontan physiology was created
in order to determine the appropriate boundary conditions
for the simulation of a perfect Fontan, moderate
Fontan dysfunction, and Fontan failure. We
further experimentally evaluated the suction potential
of the axial flow blood pump using a physical
elastomer material for the IVC.
MATERIALS AND METHODS
Latest pump design
Figure 1 illustrates the latest design of the intravascular
axial flow blood pump, which is designed for
percutaneous positioning in the IVC or extracardiac
conduit. The outer protective cage has radially
arranged filaments that serve as touchdown surfaces
to protect the vessel wall from the rotating
components.The rotating pump consists of an impeller
with three uniquely designed blades, having
appropriate angles to achieve the desired operating
range. In addition, the pump has a stationary region
of diffuser blades that are located on the protective
cage. These diffuser blades convert the rotational
energy by the impeller to potential energy or pressure
through a shift in flow directionality or
straightening. Pump rotation is induced through a
motor-magnetic bearing suspension, which levitates
and rotates the impeller within the protective cage.
Idealized and patient-specific cavopulmonary
circulation
To assess the interactive dynamics between the
pump and the Fontan physiology, four numerical
models were constructed, as shown in Fig. 2, using the
computer-aided design (CAD) software SolidWorks
(SolidWorks, Concord, MA, USA). The models consisted
of (i) an idealized TCPC with vessel diameters
of 18 mm and having a standard one-diameter offset
to establish the cavopulmonary junction; (ii) an idealized
TCPC of similar geometric characteristics
having a blood pump in the IVC; (iii) an anatomical
model of the cavopulmonary circulation in a 13-yearold
Fontan patient; and (iv) the same anatomical
model having a blood pump in the IVC.
After investigational review board approval (IRB
#HM11360), we performed a retrospective study of
patient data and harvested two-dimensional (2D)
magnetic resonance imaging (MRI) data of an extracardiac
Fontan circulation. The patient-specific anatomical
model was created by transforming the 2D
MRI into a three-dimensional (3D) CAD solid body.
We employed the software Mimics (Materialise,
Leuven, Belgium) to generate a 3D point cloud mesh
from the patient’s MRI images. The 3D point cloud
underwent smoothing iterations before importing
into SolidWorks in order to build the solid body
FIG. 1. Intravascular axial flow blood pump for Fontan patients.
Design consists of a catheter, protective cage of filaments, impeller
blade set, and diffuser blade set.
2 A.L. THROCKMORTON ET AL.
Artif Organs, Vol. ••, No. ••, 2011
model. Connecting vessel sections of the TCPC were
extended for computational fluid dynamics (CFD)
analysis with tapering of vessels to ellipse forms.
Placement of the blood pump was in the IVC for both
models of mechanical cavopulmonary assistance.
Figure 3 illustrates the patient-specific TCPC model
generation.
Computational analyses
Similar to prior numerical studies (4–7),we utilized
ANSYS CFX 12.1 software (ANSYS Incorporated,
Canonsburg, PA, USA) to simulate the fluid dynamics
through the pump and Fontan physiological
models. The program CFX-Mesh (ANSYS) was
employed to generate the tetrahedral elements. The
FIG. 2. Four numerical models as constructed
for this study. (A) Idealized TCPC
with vessel diameters of 16 mm and having
a standard one-diameter offset to establish
the cavopulmonary junction. (B) Anatomical
model of the cavopulmonary circulation in a
13-year-old Fontan patient. (C) Idealized
TCPC of similar geometrical characteristics
having a blood pump in the IVC. (D) Anatomical
model having a blood pump in the
IVC. IVC, inferior vena cava; LPA, left pulmonary
artery; RPA, right pulmonary artery;
SVC, superior vena cava; TCPC, total
cavopulmonary connection.
FIG. 3. Generation of the patient-specific TCPC model. (A) Smooth point cloud mesh imported into SolidWorks. (B) Surface knit to solid
body. (C) Vascular extensions to the solid body TCPC. MRI, magnetic resonance imaging; TCPC, total cavopulmonary connection; 2D,
two-dimensional; 3D, three-dimensional.
PATIENT-SPECIFIC MODELING OF CAVOPULMONARY ASSIST 3
Artif Organs, Vol. ••, No. ••, 2011
mesh density for each model exceeded 2.6 million
elements. Upon completion of mesh generation, the
computational flow model was implemented in the
Reynolds-averaged Navier–Stokes fluid solver,
ANSYS CFX 12.1. The k-e turbulence model was
selected based on the successful correlation of the
bulk performance (i.e., pressure-flow characteristics)
in geometrically similar prototypes from previous
work (4–6). A grid density and a convergence study
were completed for mesh quality assurance, where
incremental adjustments to the grid size were made
until the performance results deviated less than 2%.
Boundary conditions
Under steady flow conditions, we applied the
no-slip boundary condition to the stationary walls of
the models. In the stationary reference frame, the
TCPC configuration (whether idealized or patientspecific),
catheter, cage filaments, and diffuser blade
surfaces were defined as stationary boundaries. The
pump rotor, however, was specified to be in the rotating
reference frame with rotation in the counterclockwise
direction in accordance with the impeller
blade orientation. The frozen rotor interface connected
regions of differing reference frames and
maintained flow properties without circumferential
averaging. A uniform mass inflow rate or cardiac
output was specified for each simulation based on a
64%/36% flow split between the IVC and superior
vena cava (SVC), as appropriate for this Fontan
patient. The pump rotational speeds were evaluated
at 2000, 3000, 4000, 5000, and 6000 rpm. The outlet
boundary conditions, such as the left and right pulmonary
arteries (LPA and RPA), were defined to
have static and equal pressures of 10, 14, 18, 22, and
26 mm Hg. All of the vessel walls for the IVC, SVC,
and pulmonary arteries are modeled as rigid tubes.A
constant viscosity value of 0.0035 kg/m ¥ s and fluid
density of 1050 kg/m3 were used. Correspondingly,
the lumped parameter boundary conditions were
employed for the three physiologic cases and their
subsequent simulations.
Blood damage estimation
We performed a blood damage analysis on all of
the CFD models to examine the potential for
hemolysis and thrombosis. This damage model has
been widely employed as a predictive tool in the
development of several rotary blood pumps (6–8).
This approach calculates a scalar stress (s), including
the six components of the stress tensor, which
represents the level of stress experienced by the
blood (9),
s = ( S(s – s ) +Ss ) 1
6
2 2
1 2
ii jj ij
/
(1)
Past designs have considered a maximum stress
value of 425 Pa for 600 ms as the design criterion
(10). As part of this analysis, we examined fluid
streamlines as indicative of predicted fluid residence
times within the models. Using a power law relationship
between the scalar stress level and the exposure
time, a blood damage index was estimated for the
selected models (7). The accumulation of stress and
exposure time was added along the streamlines. This
approach provides a statistical estimate of damage to
blood cells traveling through this blood pump,
according to the following power law equation:
D t
inlet
outlet
= S 1.8×10-6 ·s1.991 ·? 0.765 (2)
where D represents the blood damage index and
indicates a “probability” of damage to red blood cells,
s corresponds to the scalar stress, t corresponds to the
stress exposure time, and inlet and outlet symbolize
the entrance and exit faces in the CFD model,
respectively. The inlet domains represent the SVC
and IVC, while the outlet domains are the LPA and
RPA. The numerical constants in Eq. 2, relating the
stress to the exposure time, were obtained by regression
of experimental data in a Couette viscometer
with an exposure time of 0.0034–0.6 s for fluid
stresses of 40–700 Pa (7,11). This range of investigation
is comparable with the flow conditions and stress
levels found in blood pumps. The blood damage
index is represented as a percentage of the change in
hemoglobin levels due to blood trauma divided by
the original hemoglobin content. We seek a blood
damage index below 2% for our target design (12).
Fontan cardiovascular model
An electrical analog was created to describe a
lumped parameter model of the Fontan physiology
and to generate more accurate boundary conditions
for the numerical analyses. Figure 4 displays the
lumped parameter model. In this model, resistors correspond
to the viscous property of blood flow, and
inductances represent the inertial effects of blood
flow. The capacitors model the elastic properties or
compliance of the vessel walls, and the diodes are
used to mimic the properties of one-directional
valves. The volumetric flow rate correlates to the
current flow in the electrical analog. It consists of
analog blocks for the single ventricle, aorta and systemic
arterial bed, peripheral circulation, systemic
4 A.L. THROCKMORTON ET AL.
Artif Organs, Vol. ••, No. ••, 2011
venous circulation (IVC and SVC), pulmonary arteries,
and pulmonary veins. The time-varying elastance
and contraction properties of the single ventricle
were based on previous work, according to reference
(13). The parameter values for this model are shown
in Table 1 (13–18). Using Matlab (MathWorks,
Natick, MA, USA) this Fontan cardiovascular model
provided boundary conditions for three patient conditions:
(i) perfect Fontan physiology; (ii) moderate
dysfunction; and (iii) Fontan failure. The perfect
Fontan physiology is representative of an older child
who received the Fontan completion and has stable
cavopulmonary hemodynamics, whereas the moderate
dysfunction case corresponds to conditions where
the caval and pulmonary arterial pressures are on the
rise. The Fontan failure condition signifies exceedingly
high caval and pulmonary arterial pressures
with compromise of ventricular function. Table 2 lists
FIG. 4. Electrical circuit analog of a Fontan circulation (hypolastic left heart syndrome). RA, capacitance of the right atrium; RRA,
resistance of mitral valve; DRA, mitral valve; SV, capacitance of systemic ventricle; RSV, resistance of systemic ventricle; Rshunt, atrial septal
connection from left atrium to right atrium; LA, capacitance of the left atrium; DAO, aortic valve; RAO, aortic valve resistance; LAO, inductance
of blood in the aorta; CAA, capacitance of the ascending aorta; RAA, resistance of the ascending aorta; LAA, inductance of blood in
ascending aorta; CSAB, capacitance of the systemic arterial bed; RSAB, resistance of the systemic arterial bed; LSAB, inductance of blood
in the systemic arterial bed; CSVC, capacitance of the superior vena cava (SVC); CIVC, capacitance of the inferior vena cava (IVC); RSVC,
resistance of the SVC; RIVC, resistance of the IVC; CPA, capacitance of the pulmonary artery; LPA, inductance of the pulmonary artery; RPA,
resistance of the pulmonary artery; CPAB, capacitance of the pulmonary arterial bed; RPAB, resistance of the pulmonary arterial bed; CPVB,
capacitance of the pulmonary venous bed; and RPVB, resistance of the pulmonary venous bed.
TABLE 1. Parameter values for electrical analog model of Fontan physiological
states
Parameter
Analog parameter value
Perfect Fontan Moderate dysfunction Fontan failure
ESV 3.1 mm Hg/mL 2.8 mm Hg/mL 2.2 mm Hg/mL
RLA, RRA 0.004 mm Hg s/mL 0.004 mm Hg s/mL 0.004 mm Hg s/mL
RSV 0.12 mm Hg s/mL 0.12 mm Hg s/mL 0.12 mm Hg s/mL
RAO 0.004 mm Hg s/mL 0.004 mm Hg s/mL 0.004 mm Hg s/mL
LAO 0.01 mm Hg s2/mL 0.01 mm Hg s2/mL 0.01 mm Hg s2/mL
CAA 0.12 mL/mm Hg 0.12 mL/mm Hg 0.12 mL/mm Hg
RAA 0.8 mm Hg s/mL 0.8 mm Hg s/mL 0.8 mm Hg s/mL
LAA 0.01 mm Hg s2/mL 0.01 mm Hg s2/mL 0.01 mm Hg s2/mL
CSAB 1.0 mL/mm Hg 1.0 mL/mm Hg 1.0 mL/mm Hg
RSAB 0.25 mm Hg s/mL 0.4 mm Hg s/mL 0.96 mm Hg s/mL
LSAB 0.006 mmHg s2/mL 0.006 mm Hg s2/mL 0.006 mm Hg s2/mL
CSVC 3.0 mL/mm Hg 3.0 mL/mm Hg 3.0 mL/mm Hg
RSVC 0.05 mm Hg s/mL 0.058 mm Hg s/mL 0.077 mm Hg s/mL
CIVC 3.0 mL/mm Hg 3.0 mL/mm Hg 3.0 mL/mm Hg
RIVC 0.028 mm Hg s/mL 0.030 mm Hg s/mL 0.034 mm Hg s/mL
CPA 4.1 mL/mm Hg 4.1 mL/mm Hg 4.1 mL/mm Hg
LPA 0.001 mmHg s2/mL 0.001 mm Hg s2/mL 0.001 mm Hg s2/mL
CPAB 2.5 mL/mm Hg 2.5 mL/mm Hg 2.5 mL/mm Hg
RPAB 0.11 mm Hg s/mL 0.18 mm Hg s/mL 0.25 mm Hg s/mL
CPVB 13 mL/mm Hg 8 mL/mm Hg 4.2 mL/mm Hg
RPVB 0.068 mm Hg s/mL 0.068 mm Hg s/mL 0.068 mm Hg s/mL
HR 78 beats/min 85 beats/min 100 beats/min
**Parameters not directly listed or defined have a negligible impact as compared with the
defined parameters.
HR, heart rate.
PATIENT-SPECIFIC MODELING OF CAVOPULMONARY ASSIST 5
Artif Organs, Vol. ••, No. ••, 2011
the boundary conditions as implemented in the
patient-specific numerical models.
Energy calculations
To assess the impact of the blood pump in the IVC
on the total energy of the cavopulmonary flow conditions,
we used a simplified control volume approach
to calculate the energy losses through TCPC configuration
with and without the pump (19).As a common
approach used when considering surgical optimization
of the TCPC, this analysis allowed for the
estimation of the energy loss or gain in the cavopulmonary
configuration, according to the following
equation:
E P u u u n A
P Q P
loss static k k i i i
total_in inlet total_
= – +
= ( ) –
S( 0.5? )
S S( out )Qoutlet
(3)
where
Ptotal = Pstatic + 0.5?uiui (4)
Qi = uiAi (5)
r corresponds to the fluid density, Pstatic is the static
fluid pressure, ui symbolizes the components of the
velocity vector, ni symbolizes the components of the
outward surface normal vector of the control surfaces,
Eloss represents the rate of energy consumption
within the control volume, Ptotal is the total pressure
including the static pressure component in addition
to the kinetic energy component, and Qi is the flow
rate at an inlet or outlet.The inlet domains represent
the SVC and IVC, while the outlet domains are the
LPA and RPA.
Suction experimental apparatus and measurements
As this intravascular pump will be placed in the
extracardiac conduit or IVC, we performed an initial
suction limitation study. Figure 5 illustrates the
experimental setup.This test evaluated the degree of
suction, if any, of the pump using 1/16?, 3/32?, and 1/8?
thick Sylgard 184 elastomer tubing,which is a material
that is routinely used to model the compliance of the
venous vessels in mock circulatory loops (20). Details
of the test loop and the general experimental protocol
can be found in the literature (21). Considering a
worst case,we tested the impeller without a protective
cage to support the surrounding tubing and performed
the measurements at a higher operating speed
of 9000 rpm than would be normally employed. Using
high-precision calipers,we measured the deformation
on both sides and the top of the 6? in-length tubing at
different axial lengths along the tubing, moving closer
toward the pump.The flow rate was set to 3.5 L/min as
the operating design point.These measurements were
repeated five times for each tubing thickness, and
average values are reported.
RESULTS
Pump characteristics
Figure 6 shows the numerical findings for the pressure
generation of the Fontan pump for 2000–
6000 rpm.The pump was able to deliver 1–16 mm Hg
over of the flow range of 0.5–4.5 L/min for those
operating speeds.Minor losses in pressure generation
were observed at the highest flow rates, as would be
expected. Otherwise, the pressure generation was
maintained and consistent for each rotational speed.
To examine the pump and its effect on the IVC
pressure (i.e., pump inlet), we performed simulations
using the patient-specific, anatomical model to
predict the IVC pressure as a function of increasing
rpm, as shown in Fig. 7. These data correspond to an
operating condition with LPA and RPA pressures
ranging from 10 to 26 mm Hg and a design flow rate
TABLE 2. Results of the numerical simulations for the Fontan physiological states and the inclusion of the pump in the
inferior vena cava
Parameter or boundary conditions
Perfect Fontan Moderate dysfunction Fontan failure
No pump Pump No pump Pump No pump Pump
PLPA (mm Hg) 8.73 8.73 12.92 12.92 20.35 20.35
PRPA (mm Hg) 8.73 8.73 12.92 12.92 20.35 20.35
PIVC (mm Hg)** 9.08 1.26 13.26 5.47 20.49 12.57
PSVC (mm Hg) 9.10 9.24 13.27 13.41 20.54 20.56
QLPA (L/min) 1.86 1.93 1.84 1.92 1.02 1.09
QRPA (L/min) 2.15 2.08 2.14 2.06 1.21 1.41
QIVC (L/min) 2.55 2.55 2.30 2.30 1.50 1.50
QSVC (L/min) 1.46 1.46 1.68 1.68 0.73 0.73
Pump (rpm) 0 4000 0 4000 0 4000
Energy gain (mW) -2.5 40.7 -2.3 36.4 -0.7 25.6
**Location at the pump inlet.
6 A.L. THROCKMORTON ET AL.
Artif Organs, Vol. ••, No. ••, 2011
of 3.5 L/min. Figure 7 illustrates that the IVC pressure
(i.e., pump inlet) was found to decrease as the
pump was rotated at higher speeds. Higher arterial
pressures led to elevated IVC pressures.These results
also indicated limitations for suction potential for
certain operating situations in which cavopulmonary
pressures are at more normal physiologic levels.
Blood damage findings
The blood damage analysis was conducted for all
four models and included the release of 800 inert
particles in the IVC and SVC. Table 3 describes the
blood damage analysis findings. Figure 8 displays the
fluid streamlines of the patient-specific and pumpsupported
case, where a rotational component of the
flow is observed entering the TCPC junction and
resolving itself in the larger pulmonary arteries. The
maximum particle residence times for the pumpsupported
cases were considerably lower at 0.6
(ideal) and 0.9 s (anatomic) in comparison with 2.4–
2.5 s for the nonsupported cases. Maximum scalar
stresses were approximately 3 Pa for both the idealized
and anatomic models without mechanical
assistance. The maximum scalar stress values in the
pump were found to be only 290 Pa, which is much
lower than the 425 Pa upper limit. For the non-pumpsupported
models, the mean values of the blood
damage index were determined to be 1.3 ¥ 10-4%
(ideal) and 8.9 ¥ 10-6%(anatomic).The mechanically
assisted cases resulted in higher mean blood damage
indices of 0.3% (ideal) and 0.4% (anatomic) with
FIG. 5. Suction limitation studies with the Sylgard tubing. (A) Hydraulic flow loop consisting of an inlet and outlet tank, the pump housing
in between the two reservoirs, the pump prototype, pressure transducer, pressure signal conditioner, flow probe and meter, power source
for the motor, motor controller, motor, drive shaft and bearing support system, and Labjack data acquisition A/D board. (B) The Sylgard
tubing tested with different sizes as evaluated. (C) Location of tubing in between the inlet and outlet tank with the pump retracted to clearly
show the tubing. Axial locations where the measurements were taken are shown. BLDC, brushless direct current.
FIG. 6. Numerical predictions of the pressure generation of the
axial flow blood pump for rotational speeds of 2000–6000 rpm.
FIG. 7. CFD predictions of the IVC pressure (pump inlet side) as
a function of increasing rotational speed. The impact of the blood
pump and its effect on the IVC pressure on the inlet side of the
pump as a function of increasing rotational speed and pulmonary
arterial pressures is observed. The IVC pressure is decreased as
the pump is rotated faster, as would be expected. Additionally,
higher arterial pressures lead to higher-IVC pressures. CFD,
computational fluid dynamics; IVC, inferior vena cava.
PATIENT-SPECIFIC MODELING OF CAVOPULMONARY ASSIST 7
Artif Organs, Vol. ••, No. ••, 2011
maximum indices of 0.7% (ideal) and 1.5%
(anatomic).
Hydraulic energy estimations
Figure 9 demonstrates the energy gain due to
mechanical assistance of the idealized and anatomical
TCPC configuration in comparison with the no
pump support conditions. Figure 9 illustrates the
results for the energy gain as a function of increasing
flow rate, and Fig. 10 shows the findings for the
energy gain in the anatomical model of the TCPC as
a function of increasing rpm. In Fig. 9, these simulations
occurred at an operating condition with LPA
and RPA pressures of 14 mm Hg, a range of flow
rates from 0.5 to 4.5 L/min, and a rotational speed of
4000 rpm. In contrast to the energy loss associated
with the idealized and anatomical TCPC without
mechanical support, the CFD predictions indicated
that the use of the pump to mechanically augment
pressure in the IVC increases the hydraulic energy of
the TCPC as a function of flow rate and rotational
speed.
Cardiovascular modeling of Fontan physiological
states
Table 2 describes the boundary conditions, as produced
by the cardiovascular model, for the three different
physiologic states of the Fontan physiology
and the CFD results for the simulations. This table
TABLE 3. Blood damage analysis results for each CFD model
Analysis characteristic
No pump support Pump support
Idealized TCPC Patient TCPC Idealized TCPC Patient TCPC
Mean particle residence time (s) 0.68 0.68 0.6 0.9
Time for 75% of particles to exit (s) 0.92 1.0 0.3 0.3
Maximum scalar stress (Pa) 2.8 3.0 290 290
Mean blood damage index (%) 1.3 ¥ 10-4 8.9 ¥ 10-6 0.3 0.4
CFD, computational fluid dynamics; TCPC, total cavopulmonary connection.
FIG. 8. Comparison of fluid streamlines during mechanical
cavopulmonary assistance for the patient-specific model.
FIG. 9. Energy gain due to mechanical assistance of the TCPC
with a blood pump in the IVC. Mechanical assistance of the IVC
pressure enhanced the hydraulic energy within the TCPC as
compared with conditions without pump support. Left and right
pulmonary arterial pressures are equal at 14 mm Hg, and the
rotational speed of the pump is 4000 rpm. IVC, inferior vena
cava; TCPC, total cavopulmonary connection.
FIG. 10. Energy gain in the patient-specific anatomical model
having mechanical assistance in the inferior vena cava with
increasing pump rotational speed. Left and right pulmonary arterial
pressures are equal at 14 mm Hg.
8 A.L. THROCKMORTON ET AL.
Artif Organs, Vol. ••, No. ••, 2011
also presents the results of this portion of the study
for the pump operating at a rotational speed of
4000 rpm. The pump was able to successfully
augment the hydraulic energy of the TCPC junction
for all of the physiologic conditions. Each simulation
allowed for the opportunity to evaluate the inlet
pressure of the pump; the inlet pressure of the pump
never dropped below 0 or showed evidence of
suction conditions based on the simulations. As is
also clinically observed, the flow rate in the RPA was
found to exceed that in the LPA, which is largely due
to the IVC vessel offset toward the RPA.
Suction experiments
Tables 4 and 5 display the vertical and horizontal
deformation measurements, respectively, for the
suction limitation study for the Sylgard tubing. The
variation in the vertical deformation was found to not
exceed 2.8% as compared with the initial dimension
measurements at the left, center, and right locations.
Similarly, the horizontal deformation measurement
deviation at each location was determined not to
exceed 2.2%. These results indicate minimal deformation
to the Sylgard tubing during the suction study,
and data trends demonstrated that the tubing, regardless<